Self-feature Learning

Abstract

Deep learning-based models have achieved unprecedented performance in single image super-resolution (SISR). However, existing deep learning-based models usually require high computational complexity to generate high-quality images, which limits their applications in edge devices, e.g., mobile phones. To address this issue, we propose a dynamic, channel-agnostic filtering method in this paper. The proposed method not only adaptively generates convolutional kernels based on the local information of each position, but also can significantly reduce the cost of computing the interchannel redundancy. Based on this, we further propose a simple, yet effective, deep lightweight model for SISR. Experiment results show that our proposed model outperforms other state-of-the-art deep lightweight SISR models, leading to the best trade-off between the performance and the number of model parameters. CCS CONCEPTS• Computing methodologies → Image processing; Computational photography.

Dynamic spatial aggregation network for joint denoising and HDR imaging

“…Additionally, Oh et al 6 incorporated user priors regarding moving objects into the low-rank model and improved the performance. Inspired by the successes of deep learning models in numerous vision tasks, [7][8][9][10][11][12][13][14][15][16][17][18][19][20][21][22][23][24][25][26] Yan et al 27 integrated spatial attention mechanisms into deep networks, which effectively mitigate misaligned content during HDR image reconstruction. However, motion removal-based methods, particularly in the presence of large-scale object motions in LDR images, tend to exclude a considerable number of pixels before merging the input LDR images.…”